Recent updates to the website, services and customer dashboard
Transformer Lab is a new native tool to experiment with large AI models. Works with Apple MLX. Main features:
- Download LLMs
- Chat with LLMs
- Calculate embeddings
- Create and download datasets for training LLMs
- Finetune and Train LLMs
On many pages, you will now see a new comments section to vote on the page topic or to add your own comments. MacWeb website is more interactive than ever.
For enhanced security and a smoother experience, you must now connect to your account before placing an order.
Due to exceptionally high demand, numerous on-demand configurations are currently unavailable. We are actively working to expand the availability of computers in our data centers. Please visit the configurations section regularly for updates on availability, or consider requesting a quote. Priority is currently given to volume orders with a term commitment.
Our community section is now fully open with feature requests and general discussions
Most Linux editions are facing a major security issue with the recent discovery of XZ backdoor with SSH protocol. Apple silicon Macs are not impacted but it's better to run “brew update” if you are using Homebrew packages manager.
Brewer X is a great companion for developers. Brewer X is an Homebrew client native app for macOS. Discover new apps, update everything with single click, and automate maintenance tasks. A gem of the Mac world.
Step 1. Download Jan app from your Mac server and your VNC connection.
Step 2. Download the AI model of your choice from the app
Step 3. Enable the server with all options (coding icon).
That's it. For any native app or web app using api.openai.com API's, you can now replace the address with the the IP address of your Mac or custom domain. If you need assistance, we can help. Jan is more user friendly than LMStudio that we also recommended.
We have a new AI assistant trained on the content of the website. The assistant can reply to most questions. For human assistance, you can still call us or send a message per email.
Version 2.1 of MacWeb Computing is available. Our free utility allows you to access your Macs in the cloud directly from the menubar of macOS.
Many developers are working on Apple MLX framework that is fully open source. It's fully optimized for Python, C and Swift languages. MLX is getting traction thanks to the incredible performance. We created a new page to showcase the main advantages. The community also created a new open source project named MLX server. Inference made easy!
We recommend LMStudio app as an easy AI inference server.
Step 1. Download LMStudio app from your Mac server and your VNC connection.
Step 2. Download the AI model of your choice from the app
Step 3. Enable the server with all options (double arrows icon).
That's it. For any native app or web app using api.openai.com API's, you can now replace the address with the the IP address of your Mac or custom domain. If you need assistance, we can help.
We are working with CirrusLabs to offer VM cluster configurations with TART and Orchard engines that we are recommending.
Contact us if you have CI, VM or any cluster projects.
This native app for macOS is allowing developers to switch between multiple versions of Xcode.
https://www.xcodes.app
MacWeb Computing is now recommending configurations for developers. It's the best time to build innovative apps! We organized it in three categories.
Solutions for native apps
Solutions for web apps
Solutions for AI
We have a new page where you can compare the specs of all Mac configurations available in the cloud.
https://www.macweb.com/compare
For all Mac cloud configurations, it's possible to use a native version of Linux named Fedora Asahi. https://asahilinux.org/
It's fully optimized for Apple silicon performance. We can install it for you and setup a VNC SSH connection, default boot too.
Contact us if you are interested by this option.
Dear Valued Developers, Happy New Year! As we step into 2024, we extend our warmest wishes to you, our esteemed developer community. This year promises to be an exciting one for all of us at MacWeb Computing. We're committed to making 2024 a year of unparalleled improvements and innovations in our cloud services. Our team is working tirelessly to enhance your development experience, ensuring that our service is not only more robust and efficient but also more intuitive and developer-friendly. We're excited to support your incredible work. Thank you for choosing us as your cloud service partner. Here's to a successful and groundbreaking 2024!
From today and for all new orders, macOS Sonoma 14.2 will be the default operating system version for the best user experience and a stronger security. You can still request custom configurations with older versions of macOS from our quotes form. Remote login (SSH)is now also activated by default for easy reboot of your computer from your dashboard. You can disable it at any time from a VNC connection in macOS preferences.
Two times faster than Pytorch, fully optimized for Apple silicon architecture and unified memory, the MLX framework made by Apple is impressive. It's fully open source. We updated the Inference page.
We updated the page showcasing the best solutions for PHP development on macOS. https://www.macweb.com/php
With LM Studio, you can run LLMs on your Mac server, use models through the in-app Chat UI or an OpenAI compatible server, download any compatible model files from Hugging Face repositories, discover new & noteworthy LLMs in the app's home page. A great developer tool fully optimized for Apple silicon. https://lmstudio.ai
Sanctum is more simple to use with a beautiful user interface and a focus on local assistants. https://sanctum.ai
For running a cluster of Mac computers, Kubernetes is one of the best solution, fully optimized for Apple silicon. Minikube is helping the setup and management. https://minikube.sigs.k8s.io/docs/
To better understand the performance of each cloud instance, we added the best XcodeBenchmark for Xcode 14. Version 15 is a bit faster but most developers are still using version 14.
Cilicon is a macOS App that leverages Apple's Virtualization Framework to create, provision and run ephemeral CI VMs with near-native performance. It currently supports Github Actions, Buildkite Agent, GitLab Runner and arbitrary scripts. Depending on your setup, should be able to get up and running with your self-hosted CI in minutes. https://github.com/traderepublic/Cilicon
On the self-hosted page, we are presenting professional open source or commercial solutions for developers and businesses that can be hosted on a Mac server. https://www.macweb.com/self_hosted
As a reference point, we added the best Geekbench 6 score for single core and multi core to the list of configurations available on-demand. https://www.macweb.com/macmini
LlamaGPT is a self-hosted, offline, ChatGPT-like chatbot. Powered by Llama 2. 100% private, with no data leaving your server. There is even Code Llama support! It just requires Xcode and Docker to be installed. https://github.com/getumbrel/llama-gpt
MacWeb Computing team is thrilled about Mojo, a revolutionary programming language explicitly crafted for Artificial Intelligence, now seamlessly optimized for Apple Silicon. Created by the visionary Chris Lattner and developed by Modular, Mojo ushers in a new era of computational excellence on the Mac platform. Its harmonious design leverages the full extent of Apple’s advanced hardware, delivering unbelievable performance enhancements, and setting a new standard for AI deployment environments. What makes Mojo stand out is its fluid integration with the Mac ecosystem, designed from the ground up to support the nuances of AI development, ensuring developers can work with an unprecedented level of power and precision. The future of AI on Mac is here, and MacWeb Computing is at the forefront, offering our users the tools to unlock unimaginable potential, transforming vision into reality with Mojo. It’s using the same syntax than Python, and faster than plain C. Learn more on our website:
https://www.macweb.com/mojo
You can now run macOS inside macOS, build images using Docker and distribute them using registries.
https://macoscontainers.org
From your list of servers, you can force a restart of your cloud instance with one click. You just need to enable Remote Login in macOS preferences for the cloud instance. If you customized the password or added a user, you also need to edit the information of your cloud instance. A confirmation dialog is always displayed.
You can now edit your cloud instances information, add notes, add a custom user or custom password, even a custom name.
Login and password is now automatically added to the “Connect via VNC” button.
We recommend to test both solutions on a new cloud instance. We tested both solutions for months without any major issues but it's better to do personal testing if you have a sensitive workflow.
You can now browse more content on the FAQs section of our website
We recently joined Vapor community and we are happy to help developers supporting Swift on servers.
We are happy to introduce our new name MacWeb Computing, our new branding, and our new website showcasing our ambition for cloud services. Previous customer accounts are fully supported.